We investigate the ability of 1-D ground response simulations to match observed levels of site amplification from California vertical arrays. Using 10 vertical arrays, we find simulations to best match data using a VS-based damping model from the literature. We find a higher percentage of California sites, as compared to KiK-net sites from Japan, to have a reasonable match of empirical and theoretical transfer function shapes. The empirical transfer functions also have a greater degree of event-to-event consistency than has been found previously in Japan. Cases with poor matches highlight that 1-D simulations can fail to accurately model site response.

Site-specific geotechnical ground response analyses (GRAs) are typically performed to evaluate stress and strain demands within soil profiles and/or to improve the estimation of site response relative to generic site terms from empirical prediction equations. Implementation of GRA results in probabilistic seismic hazard analysis (PSHA) requires knowledge of the mean and standard deviation of site amplification from GRA. We provide expressions for evaluating within-event standard deviations of ground motion given the parameterization of site amplification for a particular intensity measure into a nonlinear equation. A key component of those expressions is the standard deviation of site amplification, φlnY. We evaluate from the literature estimates of φlnY from ground response simulations using variable inputs and from interpretations of ground motion recordings. We find data-based studies to exhibit relatively consistent values of φlnY ≈ 0.3 over a wide range of oscillator periods, whereas simulation-based studies produce much higher values below the fundamental period of the analyzed soil column and lower values at longer periods. Values derived from the data-based studies are recommended for application.